This paper analyzes a model in which housing tenure choice serves as a means of screening households with different utilization rates. If the proportion of low-utilization types is small, there is a separating equilibrium at which tenure choice acts as a screening device: consistent with empirical evidence, low-utilization households buy a house, while high-utilization types rent. Otherwise, there is a pooling equilibrium. The reason why, contrary to standard screening models, a pooling equilibrium possibly exists is indivisibility of home ownership, which makes it a very costly screening device. Introducing partial ownership restores the standard results: non-existence of a pooling equilibrium and possible non-existence of equilibrium.